News • Study on devices and implants

AI could improve speech recognition in hearing aids

In noisy environments, it is difficult for hearing aid or hearing implant users to understand their conversational partner because current audio processors still have difficulty focusing on specific sound sources. In a feasibility study, researchers from the Hearing Research Laboratory at the University of Bern and the Inselspital are now suggesting that artificial intelligence (AI) could solve this problem.

golden ear-shaped art installation on concrete wall

Image source: Unsplash/Jaee Kim

The study has been published in the journal Hearing Research.

Hearing aids or hearing implants are currently not very good at selectively filtering specific speech from many sound sources for the wearer – a natural ability of the human brain and sense of hearing known in audiology as the “cocktail party effect”. Accordingly, it is difficult for hearing aid users to follow a conversation in a noisy environment. Researchers at the Hearing Research Laboratory of the ARTORG Center, University of Bern, and Inselspital have now devised an unusual approach to improve hearing aids in this respect: virtual auxiliary microphones whose signals are calculated by an AI.

portrait of tim fischer
Dr Tim Fischer

The more microphones are available and the more widely they are distributed, the better a hearing aid can focus on sound from a particular direction. Most hearing aids have two microphones close together due to lack of space. In the first part of the study, the Hearing Research Laboratory (HRL) determined that the optimal microphone location (for better focusing) is in the middle of the forehead – though this is a very impractical location. “We wanted to get around this problem by adding a virtual microphone to the audio processor using artificial intelligence,” said Tim Fischer, a postdoctoral researcher at HRL, explaining this unconventional approach.

For the study setup, ARTORG Center engineers used the “Bern Cocktail Party Dataset”, a collection of a variety of noise scenarios with multiple sound sources from multi-microphone recordings of hearing aid or cochlear implant users. Using 65 hours of audio recordings (more than 78,000 audio files), they trained a neural network to refine a commonly used directionality algorithm (beamformer). For improved speech understanding the deep learning approach calculated additional virtual microphone signals from the audio data mixture. 20 subjects tested the AI-enhanced hearing in a subjective hearing test accompanied by objective measurements. Particularly in cocktail party settings, the virtually sampled microphone signals significantly improved the speech quality. Hearing aid and cochlear implant users could therefore benefit from the presented approach, especially in noisy environments.

Placement of the 16 microphones used for cocktail party scenario recordings
Placement of the 16 microphones used for 'cocktail party scenario' recordings

Image source: Fischer et al., Hearing Research 2021 (CC BY-NC-ND 4.0)

"I think that artificial intelligence represents an important contribution to the next generation of hearing prostheses, as it has great potential for improving speech understanding, especially in difficult listening situations," says Marco Caversaccio, Chief Physician and ENT Department Head. As auditory assistive technologies and implants are a major focus of research at the Inselspital, important data-based foundations are being laid here for further development that should bring the natural hearing experience closer. The novel approaches will directly benefit patients within the framework of translational studies.

Although within this study the virtually added microphones significantly improved the quality of speech understanding with hearing aids, further studies still need to overcome some technical hurdles before the methodology can be used in hearing aids or cochlear implant audio processors. This includes, for example, a stable functioning directional understanding even in reverberant environments. 


Source: Inselspital, University Hospital Bern

01.07.2021

Read all latest stories

Related articles

Photo

News • Machine learning in sample analysis

Lab-trained pathology AI meets real world: ‘mistakes can happen’

AI models are highly capable in analysing tissue samples – as long as conditions are lab-perfect. Add a little contamination, however, and diagnostic accuracy goes out the window, a new study shows.

Photo

News • Finding therapy-relevant genetics

Leukaemia: AI provides support in diagnostics

Certain genetic features are crucial for treatment decisions for AML leukaemia. A team from Münster shows how an AI-based method can predict these features from images of bone marrow smears.

Photo

News • Training course and sound therapy

Can an app train the brain to overcome tinnitus?

Tinnitus is common, affecting up to one in four people. Now a study shows that the debilitating effects on mental health can often be reduced via training and sound therapy in a smartphone app.

Related products

Subscribe to Newsletter